A Somatostatin Receptor Subtype-3 (SST3) Peptide Agonist Shows Antitumor Effects in Experimental Models of Nonfunctioning Pituitary Tumors Mari C

Published OnlineFirst October 17, 2019; DOI: 10.1158/1078-0432.CCR-19-2154

CLINICAL CANCER RESEARCH | TRANSLATIONAL CANCER MECHANISMS AND THERAPY

A Somatostatin Receptor Subtype-3 (SST3) Peptide Agonist Shows Antitumor Effects in Experimental Models of Nonfunctioning Pituitary Tumors Mari C. Vazquez-Borrego 1,2,3,4, Vandana Gupta5, Alejandro Ibanez-Costa~ 1,2,3,4, Manuel D. Gahete1,2,3,4, Eva Venegas-Moreno6, Alvaro Toledano-Delgado1,3,7, David A. Cano6, Cristobal Blanco-Acevedo1,3,7, Rosa Ortega-Salas1,3,8, Miguel A. Japon 9, Ana Barrera-Martín1,3,10, Alexandre Vasiljevic11,12,13, Jason Hill5, Shengwen Zhang5, Heather Halem5, Juan Solivera1,3,7,Gerald Raverot11,12,14, María A. Galvez 1,3,10, Alfonso Soto-Moreno6, Marcelo Paez-Pereda5, Michael D. Culler5, Justo P. Castano~ 1,2,3,4, and Raul M. Luque1,2,3,4

ABSTRACT ◥ fi Purpose: Somatostatin analogues (SSA) are ef cacious and in response to SST3-agonists. Tumor growth was assessed in a safe treatments for a variety of neuroendocrine tumors, especially preclinical PitNET mouse model treated with a SST3-agonist. fi fi pituitary neuroendocrine tumors (PitNET). Their therapeutic Results: We successfully identi ed the rst SST3-agonist pep- effects are mainly mediated by somatostatin receptors SST2 and tides. SST3-agonists lowered cell viability and chromogranin-A in vitro SST5. Most SSAs, such as octreotide/lanreotide/pasireotide, are secretion, increased apoptosis , and reduced tumor growth either nonselective or activate mainly SST2. However, nonfunc- in a preclinical PitNET model. As expected, inhibition of cell fi tioning pituitary tumors (NFPTs), the most common PitNET viability in response to SST3-agonists de ned two NFPT popula- fi type, mainly express SST3 and nding peptides that activate this tions: responsive and unresponsive, wherein responsive NFPTs particular somatostatin receptor has been very challenging. expressed more SST3 than unresponsive NFPTs and exhibited a Therefore, the main objective of this study was to identify profound reduction of MAPK, PI3K-AKT/mTOR, and JAK/STAT SST3-agonists and characterize their effects on experimental signaling pathways upon SST3-agonist treatments. Concurrently, NFPT models. SSTR3 silencing increased cell viability in a subset of NFPTs. Experimental Design: Binding to SSTs and cAMP level deter- Conclusions: This study demonstrates that SST3-agonists acti- minations were used to screen a peptide library and identify SST3- vate signaling mechanisms that reduce NFPT cell viability and agonists. Key functional parameters (cell viability/caspase activ- inhibit pituitary tumor growth in experimental models that ity/chromogranin-A secretion/mRNA expression/intracellular expresses SST3, suggesting that targeting this receptor could be an signaling pathways) were assessed on NFPT primary cell cultures efficacious treatment for NFPTs.

(PitNET), where they reduce/normalize hormonal levels, shrink Introduction tumors, and improve clinical symptoms (1, 4). Most of their ther- Somatostatin receptors (SST1–5) comprise a family of seven trans- apeutic actions are assumed to be mediated through SST2 and SST5 membrane G-protein–coupled receptors able to bind and be activated activation (1). Although SSAs are able to bind other SSTs, such by somatostatin (1, 2). SST activation has been widely associated to as SST3, this binding capacity is signiﬁcantly lower in comparison multiple effects, most of them inhibitory actions on hormone secre- with SST2/SST5 (1). Remarkably, some NETs types express SST3 tion and cellular processes such as proliferation in normal and tumor at higher or similar levels than SST2 or SST5. This is the case for tissues (1). Thus, SSTs are considered as an attractive therapeutic some gastroenteropancreatic NETs (5), pheochromocytomas (6) target to treat different tumor pathologies (1, 3). Clinically available and specially nonfunctioning pituitary tumors (NFPTs), which rep- somatostatin analogues (octreotide/lanreotide/pasireotide) are used resent a heterogeneous group of tumors that constitute 30% of all to treat neuroendocrine tumors (NET), including pituitary NETs PitNETs (7–10).

1Maimonides Institute of Biomedical Research of Cordoba (IMIBIC), Cordoba, Spain. Note: Supplementary data for this article are available at Clinical Cancer 2Department of Cell Biology, Physiology and Immunology, University of Cordoba, Research Online (http://clincancerres.aacrjournals.org/). Cordoba, Spain. 3Reina Soﬁa University Hospital (HURS), Cordoba, Spain. 4CIBER Physiopathology of Obesity and Nutrition (CIBERobn), Cordoba, Spain. 5IPSEN M.C. Vazquez-Borrego and V. Gupta contributed equally to this article. 6 Bioscience, Cambridge, Massachusetts. Metabolism and Nutrition Unit, Hospital Corresponding Authors: Raul M. Luque, Maimonides Institute of Biomedical Universitario Virgen del Rocío, Instituto de Biomedicina de Sevilla (IBIS), Sevilla, Research of Cordoba (IMIBIC), Cordoba 14004, Spain. Phone: þ34-957-213-740; 7 8 Spain. Service of Neurosurgery, HURS, Cordoba, Spain. Anatomical Pathology Fax: þ34-957-213-740; E-mail: [email protected]; and Justo P. Castano,~ 9 Service, HURS, Cordoba, Spain. Department of Pathology, Hospital Universitario [email protected] Virgen del Rocío, Sevilla, Spain. 10Service of Endocrinology and Nutrition, IMIBIC, 11 HURS, Cordoba, Spain. FacultedeMedecine Lyon Est, UniversiteLyon1,Lyon, Clin Cancer Res 2020;XX:XX–XX France. 12INSERM U1052, CNRS UMR5286, Cancer Research Centre of Lyon, Lyon, France. 13Centre de Pathologie et de Biologie, Groupement Hospitalier Est, Hospices doi: 10.1158/1078-0432.CCR-19-2154 Civils de Lyon, Lyon, France. 14Fed erationd ’endocrinologie, Groupement Hospitalier Est, Hospices Civils de Lyon, Bron, France. 2019 American Association for Cancer Research.

AACRJournals.org | OF1

Vazquez-Borrego et al.

SST3-agonists through the screening of a peptide library; and (ii) Translational Relevance in vitro/in vivo determine the therapeutic potential of SST3 by fi fi fi fi Our results provided the rst successful identi cation of studying: (iiA) the direct effect of the identi ed SST3-speci c fi speci c/functional somatostatin receptor subtype-3 (SST3)- agonists/antagonists on key functional parameters for NFPTs peptidic agonists and antagonists, which demonstrated a key (i.e., cell viability/caspase activity/chromogranin-A (CgA) secre- pathologic function and therapeutic potential of SST3 in non- tion/signaling pathways/mRNA expression) and the functional functioning pituitary neuroendocrine tumor (NFPT) cells consequences of SSTR3 silencing on cell viability; and (iiB) the in vitro in vivo and . Indeed, treatment with SST3-peptidic agon- effect of a selected SST3-specific agonist on tumor growth in a ists and antagonists altered key pathophysiologic parameters in preclinical mouse model of PitNETs. NFPTs in vitro, including cell viability, apoptosis, chromogranin- A release, and expression levels of gonadotropin hormones (through the modulation of relevant signaling pathways), as well Materials and Methods as tumor growth in vivo, in a preclinical model. Importantly, the Reagents results obtained revealed two distinct populations of NFPTs All reagents used in this study were purchased from Sigma-Aldrich, fi (responsive vs. unresponsive NFPTs) in terms of inhibition of unless otherwise speci ed. Subtype selective SST3-agonists (L- aggressiveness features upon SST3-agonist treatment, wherein 796,778/BIM-355/BIM-071) and antagonists (BIM-839/BIM-152, responsive NFPTs expressed more SST3 and exhibited a pro- previously known as BN81658), were generously provided by Merck found reduction of MAPK/PI3K-AKT/mTOR/JAK/STAT sig- & Co., Inc. (L-796,778) or IPSEN Bioscience, Inc. (BIM compounds). naling pathways in response to SST3-agonists. Altogether, the Specificity of the MERCK human agonist was reported previously (20). translational research implications of these findings indicate that Detailed information of the affinity profile [expressed as Ki (nmol/L)] SST3 could have potential value as a therapeutic target in NFPTs and structure of the human SST3 agonists/antagonists is described and in tumor pathologies expressing SST3. in Fig. 1; Supplementary Table S1.

Library screening fi The identi cation of compounds that act as selective SST3-agonists Most NFPTs are silent gonadotropinomas characterized by lack of was based on the examination of IPSEN's extensive collection of hormone hypersecretion, which usually determines a delay in diag- synthetic peptides: see Supplementary Methods. nosis (10) and, therefore, are mostly detected as macroadenomas with consequently associated severe comorbidities related to mass effect Radioligand binding assays (i.e., headaches/visual defects/hypopituitarism; refs. 11, 12). Trans- The affinities of the compounds of interest for the different sphenoidal surgery is the first-line therapy and the only current human and mouse SSTs were determined by radioligand binding curative approach of NFPTs; however, it is often not definitive due assays in HEK-293 cells (ATCC) stably transfected with each K to the invasion of surrounding structures (sphenoidal/cavernous receptor. The inhibition constants ( i) were calculated from the K ¼ þ K sinuses) and tumor relapses are frequent even when resection seems following equation: i IC50/(1 L/ d) where L is the radioligand – fi K total (20% 30% of relapses at 10 years after the rst surgery; concentration and d the equilibrium dissociation constant of the refs. 13, 14). Radiotherapy has been largely used to prevent tumor radioligand for each receptor subtype. Radioligands were: [125I- relapses but long-term side effects have been reported (15, 16). To date, Tyr11]-lanreotide for SST3, [125I-Tyr]-seglitide for SST2 and fi the pharmacologic treatment options for NFPTs are insuf cient. [125I-Tyr]-BIM-23015 for SST5. Single clones were selected on Unfortunately, drugs currently available for functional PitNETs, such the basis of their mRNA expression measured by qPCR, further as dopamine agonists (DA) or SSAs, have shown poor efficacy in selected by their membrane expression of each SST, and by their NFPTs (16, 17). Moreover, in vitro treatment of NFPT primary cell response to lanreotide in cAMP assays. Membranes were isolated cultures with octreotide may not decrease cell viability (7), which could from these cells, mixed with trace concentrations of each specific be explained by the low SST2/SST5 expression levels (8, 9), as men- radioligand and a range of concentrations of the compounds of tioned above. interest, and then incubated at room temperature for 30 to 60 In this scenario, the importance of SST3 relies on the fact that SST3 minutes. Unbound radioligands were removed by filtration through has been associated with apoptotic/antiproliferative actions in various GF/C filters and membrane-bound radioactivity was measured studies using cell lines (1, 18, 19). Indeed, the apoptotic actions of using a gamma counter. Nonspecificbindingwasdefined in the m somatostatin/SSAs have been historically related to SST3 (1, 18, 19). presence of 1 mol/L lanreotide. Naturally, pasireotide has been suggested as a potential therapeutic fi option due to its ability to bind to SST3 with higher af nity than Cyclic AMP accumulation assay octreotide/lanreotide (1, 9). However, recent results from our group The functional agonist activities of the compounds of interest in demonstrate that it does not have a clear inhibitory effect on NFPT inhibiting intracellular cAMP production (stimulated by forskolin) primary cultures, being even less potent than octreotide (7), thus were also determined in HEK-293 cells stably expressing different reinforcing the notion that identification and validation of novel human and mouse SSTs. Cells were plated in 384-well plates and therapeutic approaches is necessary to manage NFPTs. Hence, cultured for 16 to 24 hours at 37C. Compounds of interest at a based on all this information, we hypothesized that a compound range of concentrations were incubated with the cells in the that preferentially binds to SST3 mightbeaneffectivetherapeutic presence of forskolin for 30 minutes. Cell homogenates were option to treat NFPTs. However, it has been very challenging to find incubated with fluorescence-labeled cAMP and anti-cAMP anti- peptides that selectively activate SST3, and its functional role and bodies for 1 hour before reading the homogeneous time resolved pathophysiologic relevance in NFPTs still need further demonstra- fluorescence. Fluorescence was measured using cAMP Dynamic 2 tion. Therefore, the aims of this study were to (i) identify specific Kit (Cisbio Bioassays).

OF2 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Role of SST3 in Nonfunctioning Pituitary Tumors

Figure 1.

Chemical structures of selected SST3-agonists and antagonist compounds. These compounds are representative of the families of peptides that showed the highest activity and selectivity for

SST3.

Patients, samples, and primary cell cultures guardian (in case of autopsy) was obtained. For in vitro assays, fresh Three different pituitary samples types have been included in the tumor samples were placed in sterile cold media and dispersed into current study: Cohort 1) Fresh samples from 80 NFPTs [mean age: single cells by mechanical and enzymatic disruption and cultured 57 (18–80) years; 42% women] and 80 functioning PitNETs [63 onto different tissue culture plates following the methods and GHomas (mean age: 43 (12–73) years; 48% women) and 17 reagents reported previously (7, 21, 22). ACTHomas (mean age: 38 (20–67) years; 94% women)] obtained during transsphenoidal surgery in different hospitals from Spain RNA isolation, reverse transcription, and analysis of gene [formalin-fixed paraffin-embedded (FFPE) pieces were also avail- expression levels by qPCR able for n ¼ 29 NFPT samples]; Cohort 2) 35 FFPE NFPTs [n ¼ 25 Information about RNA extraction, quantification, reverse tran- tumors from France (Hospices Civils de Lyon, Lyon, France) and n scription (RT) and qPCR using specific primers included in this study ¼ 10 from Spain (Hospital Universitario Virgen del Rocío, Seville, has been previously reported elsewhere by our group (22, 23). The Spain); (mean age: 63 (25–82) years; 46% women)]; and, Cohort 3) stability of the expression of three reference genes ACTB, HPRT, and 12 normal pituitary (NP) samples [mean age: 61 (44–85; 50% GAPDH was evaluated in all samples using RefFinder, a comprehen- women)] obtained during autopsies. The appropriate classification sive tool that integrates the currently available major computational of each pituitary sample collected (NP or tumor type) was con- programs (24), and found HPRT to be the most stable. Taking this into firmed by two different methods: examination by expert anatomo- account, the expression values of the genes of interest were normalized pathologists and by the molecular screening using quantitative real- to HPRT mRNA levels. time PCR (qPCR) as described previously (7, 21, 22). All techniques carried out in this study were conducted in accordance with the IHC analysis for SST3 in NFPTs ethical standards of the Helsinki Declaration, of the World Medical IHC staining for SST3 was performed using an automated immu- Association and with the approval of the University of Cordoba/ nostainer (Benchmark XT, Ventana Medical Systems) to corroborate n ¼ IMIBIC and Ethics Committees from all the Hospitals involved in the presence of SST3 in a set of 35 FFPE NFPTs [Cohort 2: 25 the study. Informed written consent from each patient or patient’ tumors from France (Hospices Civils de Lyon) and n ¼ 10 from Spain

AACRJournals.org Clin Cancer Res; 2020 OF3

Vazquez-Borrego et al.

(Hospital Universitario Virgen del Rocío)], and to evaluate the pres- the pair of duplicate spots, representing each phosphorylated site, n ¼ n ence of SST3 in SST3-agonist responsive ( 17) and unresponsive ( was calculated after subtraction of background values (pixel den- ¼ 12) NFPTs (available FFPE samples from Cohort 1). The presence of sity) from negative control spots and normalization to average antigen was revealed with the UltraView Universal DAB Detection Kit values from positive control spots using HLImageþþ sofware (Ventana). SST3 antibody was diluted 1/1000 (UMB-5, Abcam). An (Version 22.0.0a; R&D Systems). immunoreactive score (IRS) was established for each section. This score was calculated (ranging from 0 to 12) as the product of the Measurement of key genes by Western blotting – – – percentage of positive cells (0 4: 0, 10, 11 50, 51 79, and 80%, For Western blot analysis, 300,000 cells/well were cultured in – respectively) and the staining intensity (0 3: no staining, mild, mod- 12-well plates and incubated for 8 minutes with vehicle and BIM- erate, and strong, respectively). The slides were semiquantitatively 355. Proteins were extracted, separated by SDS-PAGE, and trans- scored by two independent, experienced pathologists by a double- ferred to nitrocellulose membranes(Millipore),asreportedpre- blind method and similar results were obtained. viously (26). Then, blocked membranes were incubated with the primary antibodies to detect TP53 and LHB [TP53 (Cell Signaling Measurement of cell viability in primary pituitary tumor cell Technology; #2524s; 1:1,000) and LHB (AFP-697071P; 1:1,000; cultures generously supplied by Dr. Albert Parlow; NIDDK-NIH, Bethesda, Viability of primary pituitary tumor cells was evaluated every MD)] and with appropriate secondary antibodies (anti-mouse and 24 hours (10,000 cells/well in 96-well plates) in response to SST3- anti-rabbit antibodies from Cell Signaling Technology), and devel- agonists and antagonists (treatments were daily refreshed after each oped using an enhanced chemiluminescence detection system (GE measure) using Alamar-blue reagent (Invitrogen) and a FlexStation Healthcare) with dyed molecular weight markers. A densitometric III system (Molecular Devices), as reported previously (22, 25). analysis of the bands was carried out with ImageJ software. Proteins were normalized by using total protein loading (Ponceau Measurement of caspase activity of primary pituitary tumor staining). cell cultures As a surrogate marker of apoptosis induction capacity, we measured Preclinical mouse model caspase-3/7 activity in response to different SST3-agonists and BIM- A preclinical mouse model of nonsecreting PitNET (the POMC- 839 antagonist using the Caspase-Glo 3/7 assay (Promega) according knockout generated through homologous gene targeting in embry- to the manufacturer's instructions. For this purpose, 25,000 cells/well onic stem cells) was used as reported previously (27, 28). Four weeks were plated in a 96-well white microplate and cultured for 24 hours after birth, mice were treated with vehicle or a selected SST3-agonist at 37 Cinanatmospherecontaining5%CO2. Then, cells were by subcutaneous implantation of 7-day mini-pumps (ALZET) incubated for another 24 hours with different SST3-agonists, BIM-839 for 8 weeks. Three groups were established: vehicle (n ¼ 16), low m antagonist, and vehicle. After the incubation period, 100 Lof BIM-355 dose (2.2 mg/kg/day; n ¼ 8), and high BIM-355 dose Caspase-Glo 3/7 reagent was added to each well and luminescence (7.4 mg/kg/day; n ¼ 6). Tumor volume was measured weekly by was measured at room temperature using FlexStation III system for 3 MRI imaging. After euthanasia, pituitary glands were carefully hours. fi fi removed, xed, and embedded in paraf n. IHC for SST3 and Ki67 was performed in mouse pituitary samples from vehicle-treated Measurement of CgA release in primary pituitary tumor cell animals after euthanasia (n ¼ 8). All experimental procedures cultures were carried out following the European Regulations for Animal – To evaluate the effects of SST3-agonists on CgA release, 150,000 Care, in accordance with guidelines and regulations, and under the 200,000 cells/well were used. Media were recollected after 24 hours of approval of the University/Regional's Government Research Ethics treatment and CgA concentration was evaluated using a commercial Committees. ELISA (EIA-4937; DRG). Silencing of SSTR3 gene with specific siRNAs 2þ Measurement of free cytosolic calcium concentration ([Ca ]i) To perform silencing experiments, 500,000 cells/well were plated kinetics and transfected with two specific SSTR3 siRNAs at 100 nmol/L (s13501 To assess the effects of different SST3-agonists on free cytosolic and s224690 Silencer Select Pre-designed siRNAs; Ambion) and calcium kinetics, 50,000 cells/coverslip were plated and changes in a commercial negative control (Scramble: Silencer Select Negative 2þ [Ca ]i in single cells were measured using fura-2AM (Molecular Control 1 siRNA; Ambion) using Lipofectamine RNAiMAX (Life Probes), as described previously (21, 22). Technologies). The validation of silencing conditions was carried out using the BON-1 cell line, which displays a high SST3 expression Analysis of signaling pathways by human phosphokinase (Supplementary Fig. S1). After 24 hours of transfection, cells were array detached and used for validation of the transfection efficiency (deter- To test the signaling pathways altered in responsive and unre- mined by qPCR) and for cell viability measurements as described sponsive NFPTs treated with a selected SST3-agonist, 500,000 cells/ above. To assess the SST3 protein reduction in response to the well were cultured in 12-well plates and incubated for 8 minutes silencing, 300,000–500,000 transfected cells (scramble and s13501) with vehicle and BIM-355. Specifically, three different NFPTs of were washed and lysed in SDS-DTT buffer after 48 hours of trans- each condition (vehicle/treatment; responsive/unresponsive fection. Proteins were separated by SDS-PAGE and transferred to NFPTs) were pooled and 600 mg of cell lysates were loaded in the nitrocellulose membranes (EMD Millipore). Membranes were blocked array membranes to detect relative phosphorylation levels of 43 with 5% nonfat dry milk in Tris-buffered saline with 0,05% Tween 20 relevant kinase phosphorylation sites, following the manufacturer's and incubated overnight a 4 C with primary antibodies [anti-SST3: protocol (Proteome Profiler; R&D Systems). The average signal of UMB5, ab137026 (Abcam); anti–b-tubulin: 2128S (Cell Signaling

OF4 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Role of SST3 in Nonfunctioning Pituitary Tumors

Technology)]. Secondary horseradish peroxidase–conjugated anti- SST3-agonists modulate key functional parameters in NFPT rabbit were purchased from Cell Signaling Technology. Proteins were primary cell cultures developed by using ECL detection system (GE Healthcare) with dyed SST3 was the highest expressed receptor in our cohort of NFPTs molecular weight markers. A densitometric analysis of the bands was followed by SST2, SST1, SST5, and the two SST5 splicing-variants – carried out with ImageJ software (NIH, Bethesda, MD). (SST5TMD4/SST5TMD5; Supplementary Fig. S2A C). We also found a clear differential SST3 expression pattern between functioning Statistical analysis PitNETs (GHomas/ACTHomas) and NFPTs, being this expression All data are expressed as mean SEM. Statistical differences significantly lower in GHomas/ACTHomas versus NFPTs (Supple- – between the expression levels of NFPTs and NPs were assessed by mentary Fig. S3A S3D). Moreover, SST3 IHC staining was present unpaired nonparametric Mann–Whitney tests (according to normal- in approximately 50% of samples, showing in all cases similar ity evaluated by Kolmogorov–Smirnov test). Differences between SSTs membrane and cytoplasmic localizations (Supplementary Fig. S2D), mRNA levels within NFPTs or, comparison of SST3 levels between as reported previously (9, 29). We next explored the functional role – NFPTs, functioning PitNETs, and NPs were assessed by Kruskal of SST3 using the new peptidic SST3-agonists/antagonists in com- Wallis test followed by Dunn test for multiple comparisons. All parison with the previously published nonpeptidic L-796,778 SST3- experiments were performed in a minimum of three independent agonist. All SST3-agonists induced a clear concentration-dependent primary pituitary cultures from different patients (3–4 replicates/ reduction of cell viability, being 10 7 mol/L, the lowest concentra- treatment per experiment), unless otherwise indicated. P < 0.05 was tion that caused a maximal effect, selected for further experiments considered significant. All statistical analyses were performed using (Supplementary Fig. S4A). The antagonists BIM-839 and BIM-152 GraphPad Prism 6 (GraphPad Software). See Supplementary Methods did not elicit changes on cell viability at any of the concentration for further details. tested (10 5 to 10 11 mol/L and 10 6 to 10 11 mol/L, respectively; Supplementary Fig. S4B). 7 When comparing the effects of all SST3-agonists (at 10 mol/L), we Results observed a comparable significant decrease of cell viability after 48 to Identification of SST3 peptide agonists 72 hours of incubation (23.2%, 18.3%, and 17.9% of reduction at We screened a library of synthetic peptides produced at IPSEN, 72 hours with BIM-355, BIM-071, and L-796,778, respectively), fi rst selecting for compounds that bind to SST3 and then applying a although BIM-355 seemed to be the agonist exerting a stronger multidimensional Spotfire analysis designed to identify compounds reduction (Fig. 2A). To demonstrate the specificity of these effects, with the highest selectivity for SST3 versus SST2 and SST5.Thistool we blocked them with two antagonists. Coadministration of BIM-839 was customized to display the selectivity of every compound for (10 5 mol/L) or BIM-152 (10 9 mol/L) with each agonist (10 7 mol/L) fi each of the three SSTs tested along three different axes. The SST3- fully counteracted the agonist's effect, con rming the functional role of selective compounds showed a clearly different localization in this SST3 in these effects (Fig. 2B). Remarkably, at the selected concentra- three-dimensional space clustering together on the SST3 axis. The tions, BIM-839 did not alter cell viability (Fig. 2B) but BIM-152 peptides obtained in the Sporfire analysis (i.e., BIM-23A185), while reduced cell viability at high concentrations in some tumors, which being active in receptor assays, were only hit compounds and did prompted us not to use this compound in subsequent experiments. not have optimal drug-like properties. For example, their solubility Additionally, administration of BIM-355, BIM-071, or L-796,778 and stability were suboptimal in standard quality control measure- increased caspase-3/7 activity (indicator of an increase on apoptotic ments. For this reason, small structural changes were introduced rate), BIM-355 appearing as the most potent compound. Of note, these into the structure of BIM-23A185 to obtain derivates with more effects were also completely blocked in the presence of BIM-839 drug-like characteristics. These changes yielded several peptides (Fig. 2C). (i.e., BIM-355, BIM-071, BIM-839 among others) that preserved the Furthermore, BIM-355, BIM-071, and L-796,778 clearly reduced general similarity with BIM-23A185 and had improved solubility, CgA secretion (considered as a valuable marker of NFPTs (30, 31)) stability, reproducible binding, and pharmacologic activity. Several after 24-hour incubation (Fig. 2D). Finally, we analyzed whether SST3- batches of these compounds were synthesized independently and agonists may regulate the expression of key genes related with the tested to confirm their structures and activities. High-throughput pathogenesis of PitNETs (Fig. 2E; ref. 32). Incubation with BIM-355, synthesis of these peptides allowed the characterization of their but not BIM-071, reduced gonadotropin hormone expression (FSHB binding, selectivity, and functional activity on different SSTs (Sup- and LHB), whereas BIM-071 significantly reduced the common plementary Table-S1), being BIM-23A185 discarded for further a-subunit (CGA) mRNA levels. However, incubation with L- analyses due to the low level of robustness of the results generated 796,778 did not alter CGA, FSHB,orLHB levels. Moreover, PTTG1 (data not shown). The most potent peptides (BIM-355, BIM-071, and MYC levels were not altered in response to the different agonists fi TP53 fi BIM-839) bind to SST3 with af nities in the low nanomolar range, but was signi cantly decreased in response to BIM-355 (Fig. 2E). while the binding to SST2 and SST5 is orders of magnitude weaker Validation of the downregulation in LHB and TP53 protein expression than octreotide/lanreotide/pasireotide (2). Functional cAMP assays in response to BIM-355 treatment could also be implemented by confirmed that the newly identified peptides (BIM-355, BIM-071) Western blotting in cells from a single NFPT cell culture using specific activate SST3, while BIM-839 showed SST3-antagonist activity. On antibodies and normalizing by Ponceau total protein staining the basis of these results, we selected BIM-355, BIM-071, and BIM- (Fig. 2E). 839 as the most selective and potent compounds for further Although these data demonstrate important roles of SST3 in NFPT characterization. Binding assays using human/rat/mouse SST3 cells, it should be noted that a proportion of the NFPTs analyzed did showed that BIM-071 does not activate rodent receptors. On the not respond to SST3-agonists in terms of cell viability. In all the basis of all these results, we selected these peptides to perform available NFPT cell cultures, cell viability was determined in response in vivo experiments in human NFPT primary cultures and in mice to at least two of the SST3-agonists. This analysis revealed that 16 of the with PitNETs. 42 NFPT cell cultures tested (38%) did not show a reduction of cell

AACRJournals.org Clin Cancer Res; 2020 OF5

Vazquez-Borrego et al.

Figure 2. 7 Functional assays in response to SST3-agonists and antagonists in NFPTs primary cell cultures. A, Effect of three agonists [10 mol/L: BIM-355 (n ¼ 16), BIM-071 (n ¼ 13), and L-796,778 (n ¼ 18) on cell viability in responsive tumors (24- to 72-hour treatment), measured by Alamar-blue reduction. B, Effect of BIM-355 (n ¼ 4), BIM-071 (n ¼ 3), and L-796,778 (n ¼ 9) alone or in combination with antagonists (BIM-839 and BIM-152) on cell viability. C, Effect of agonists [BIM-355 (n ¼ 5), BIM-071 and L- 796,778 (n ¼ 3)] on caspase activity (24-hour treatment), measured by Caspase-Glo 3/7 assay. D, Effect of agonists [BIM-355 (n ¼ 4), BIM-071 (n ¼ 3), and L-796,778

(n ¼ 8)] on CgA secretion (24-hour treatment), determined by commercial ELISA Kit. E, mRNA expression levels of key genes in response to SST3-agonists were measured by qPCR and adjusted by normalization factor (NF; n ¼ 7), and validation of the downregulation of LHB and TP53 at protein level was evaluated by Western blot analysis in response to BIM-355 and vehicle-treated control (n ¼ 1). Data are expressed as percent of vehicle-treated controls (set at 100%) within experiment. Values represent the mean SEM. Asterisks ( P < 0.05; P < 0.01; P < 0.001) indicate statistically signiﬁcant differences.

OF6 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Role of SST3 in Nonfunctioning Pituitary Tumors

Figure 3. Differential expression profile of key genes according to responsive or unresponsive NFPTs in terms of cell viability. A, Expression profile of somatostatin receptors in responsive (RP, black bars; n ¼ 26) compared with unresponsive tumors (URP, gray bars; n ¼ 16) and ROC curve analyses to determine the accuracy of SST3 expression as diagnostic test to discriminate between RP and URP NFPTs. Data represent median interquartile range of absolute expression levels (copy number) of each transcript adjusted by the expression level of a control gene (HPRT). B, SST3 protein levels in both responsive (RP; n ¼ 17) and unresponsive (URP; n ¼ 12) NFPTs evaluated by IHC and ROC curve analyses of SST3 expression. Data represent median interquartile range. C, Expression profile of SST3 in both RP and URP 7 tumors in response to treatment with SST3-agonists [10 mol/L; RP (n ¼ 3) and URP (n ¼ 4): BIM-355, BIM-071, and L-796,778]. Data are expressed as percent of vehicle-treated controls (set at 100%) within experiment. Values represent the mean SEM. Asterisks ( P < 0.05; P < 0.01) indicate statistically significant differences.

viability in response to any of the SST3-agonists tested and were, suggesting that only the SST3 expression pattern is able to discriminate therefore, classified as unresponsive tumors. The functional assays between responsive versus unresponsive NFPTs. Taking into account performed in each individual responsive and unresponsive NFPT and the differential SST3 expression pattern, we further investigated the the responsiveness to each SST3-agonist is shown in Supplementary capacity of different SST3-agonists to modulate this receptor expres- Tables S2 and S3. sion profile in both responsive and unresponsive tumors. This revealed that BIM-355 and L-796,778 clearly increased SST3 mRNA levels (24- Differential SST3 expression in responsive versus unresponsive hour incubation) in responsive primary NFPT cultures, whereas SST3 NFPTs expression was reduced in unresponsive NFPTs in response to BIM- To understand the differential response to SST3-agonists observed 071, but unaltered in response to BIM-355 and L-796,778 (Fig. 3C). in cell viability assays, we compared the SSTs expression levels between responsive and unresponsive tumors. Interestingly, we found higher Signaling pathways modulated in response to SST3-agonists fi P ¼ fi SST3 mRNA and protein levels (statistically signi cant changes: We next identi ed the signaling pathways underlying the ability of fi 0.029 and 0.039, respectively) in responsive versus unresponsive SST3-agonists to generate functional responses. Speci cally, NFPT fi 2þ tumors (Fig. 3A and B). However, no signi cant changes were found cells showed little response to SST3-agonists in terms of [Ca ]i in the expression of other SST-subtypes between the two NFPT kinetics (12.4%, 10%, and 7% of responsive cells to BIM-355, BIM- 2þ populations at mRNA levels (Fig. 3A). Moreover, although SST1/ 071, and L-796,778, respectively) with modest [Ca ]i reductions, SST2/SST5/SST5TMD4/SST5TMD5 mRNA levels were unable to dis- ranging between 25.6 and 27.4% (Supplementary Fig. S7A). Con- fi – tinguish between both groups, SST3 expression levels signi cantly versely, BIM-355 treated cells from responsive NFPTs showed a discriminated between responsive and unresponsive tumors (mRNA decrease in phosphorylated proteins levels involved in three major levels: AUC ¼ 0.71; P ¼ 0.03; Fig. 3A; Supplementary Fig. S5; and signaling pathways: MAPK (HSP27/PLCG1/ERK1-2/RSK1-2-3/ protein levels: AUC ¼ 0.69; P ¼ 0.098; Fig. 3B; Supplementary Fig. S6). CREB/JNK1-2-3/c-Jun), PI3K-AKT/mTOR (Akt/GSK-3a/b/Src/ We also evaluated whether any clinical parameter (extrasellar growth, FAK/p53/PRAS40/TOR/p70 S6-kinase) and JAK/STAT (STAT3/ cavernous sinus invasion, cure rate, etc.) could be different, and able to STAT5a-b/STAT6), when compared with vehicle-treated controls discriminate, between responsive and unresponsive tumors. However, (Fig. 4A and B). In contrast, BIM-355 treatment of unresponsive c2 tests did not reveal any significant result (Supplementary Table S4), NFPTs did not evoke similar reductions, and even elicited an increase

AACRJournals.org Clin Cancer Res; 2020 OF7

Vazquez-Borrego et al.

Figure 4. Measurement of intracellular signaling pathways in response to BIM-355 in NFPTs. Direct effects of BIM-355 on the phosphorylation levels of 43 kinase phosphorylation sites, measured by phosphokinase array. A, Array images are shown for responsive (left) and unresponsive tumors (right) treated with vehicle (top) or BIM-355 (bottom). A pool of three responsive and three unresponsive tumors treated with vehicle or BIM-355 were used. Each phosphoprotein array comprises two panels, which are shown in parallel. The left panel contains 28 kinase phosphorylation sites, one total protein and two reference spots, and the right panel contains 15 kinase phosphorylation sites, one total protein and one reference spot, all of them printed in duplicate. B, Graphs represent spot intensities of indicated proteins by quantifying the mean spot pixel densities. Values represent the mean SEM of duplicate spots from a pool of three responsive NFPT (NFPT 12, 14, and 24 from Supplementary Table S2) and three unresponsive NFPT (NFPT 32, 33, and 35 from Supplementary Table S2) cultures. Asterisks (, P < 0.05; , P < 0.01; , P < 0.001) indicate statistically signiﬁcant differences.

OF8 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Role of SST3 in Nonfunctioning Pituitary Tumors

of phosphorylated levels of several proteins belonging to the same Effect of SSTR3 gene silencing on cell viability pathways, such as MSK1-2/c-Jun/Src/FAK/TOR/STAT2/STAT5a-b/ SSTR3 gene silencing was achieved at the mRNA levels using two STAT5b/STAT6 (Fig. 4A and B). Interestingly, we observed clearly speciﬁc siRNAs (Fig. 5A). In addition, we could test the silencing at reduced basal phosphorylation levels in most of these proteins in protein level in two independent NFPTs at 48 or 72 hours after responsive NFPTs compared to unresponsive tumors (Supplemen- transfection, respectively. This approach revealed a SST3 content tary Fig. S7B). reduction of 70% (48 hours after transfection with s13501) and

Figure 5. Cell viability in response to SSTR3 gene silencing in NFPTs primary cell cultures. A, Validation by qPCR of SSTR3 gene silencing with two speciﬁc siRNAs [s13501 (n ¼

9) and s224690 (n ¼ 6)]. B, Effect of 24-, 48-, and 72-hour silencing of SST3 expression levels [nonresponsive (n ¼ 5) and responsive (n ¼ 4) to silencing] on cell viability, determined by Alamar-blue reduction. C, Basal SST3 expression levels in responsive (RP, n ¼ 3) compared with nonresponsive tumors (URP, n ¼ 5), before to perform the SSTR3 gene silencing. D, SST3 expression levels in responsive (RP, n ¼ 3) compared with nonresponsive tumors (URP, n ¼ 5). E, Validation by Western blotting of SST3 silencing after 48 hours of transfection using s13501 (n ¼ 1). F, Validation by Western blotting of SST3 silencing after 72 hours of transfection using s13501 and s224690 siRNA (n ¼ 1). G, Effect of 24-, 48-, and 72-hour silencing of SST3 protein alone or in combination with BIM-355 treatment on cell viability (n ¼ 1), determined by Alamar-blue reduction. Data are expressed as percent of control random siRNA (Scramble; set at 100%) within experiment. Values, mean SEM. Asterisks (, P < 0.05; , P < 0.01; , P < 0.001) indicate data that signiﬁcantly differ from scramble controls.

AACRJournals.org Clin Cancer Res; 2020 OF9

Vazquez-Borrego et al.

20% to 40% (72 hours after transfection with s13501 and s224690 with two doses of BIM-355 (group-1: 2.2 mg/kg/day and group-2: siRNA, respectively) compared with scramble-transfected cells in the 7.4 mg/kg/day). Treatment with BIM-355 reduced tumor-size after two NFPTs analyzed (Fig. 5E and F). Particularly, s13501 did not alter 8 weeks of treatment in a dose-dependent manner. This reduction cell viability, but s224690 significantly increased cell viability 48 to was statistically significantonlyatthehigherdose(Fig. 6C and D; 72 hours after transfection in 38% of NFPTs analyzed (those named it should be noted that although 6 animals started in the higher responsive to silencing) compared with scramble-transfected cells dose group, 3 of them were removed after three weeks due to an (Fig. 5B). This differential response to s224690 was not associated irritation at the minipump site. This can be clearly seen in to the initial SST3 levels (Fig. 5C), to the percentage of reduction in the Supplementary Fig. S8). level of SST3 (Fig. 5D) or to the presence of the endogenous ligands of SST3, somatostatin, or cortistatin. Indeed, this analysis revealed low levels of somatostatin (average of 0.0066 mRNA copies adjusted by Discussion HPRT) and much higher levels of cortistatin (average of 0.98 mRNA In this study, we identified for the first time two selective peptidic HPRT fi copies adjusted by ), with no signi cant differences between SST3-agonists and used them to explore the functional relevance of SSTR3 responders and nonresponders to silencing (data not shown). SST3 in NFPTs using a battery of experimental and analytical tech- fi fi On the other hand, and as expected, the treatment with BIM-355 niques. Our rst objective was to identify speci c SST3-agonists that induced a reduction on cell viability in scramble-transfected cells; could activate SST3 through the screening of a peptide library. From all fi however, this effect was completely blunted in s13501-transfected cells tested peptides, SST3-binding af nities and functional cAMP assays fi where the SST3 content was clearly reduced (Fig. 5G). This approach revealed that BIM-355 and BIM-071 were the most potent and speci c was implemented in cells from a single NFPT and suggests that the SST3-agonists, while BIM-839 was the most potent SST3-antagonist. fi effect of BIM-355 is mediated by SST3. To determine whether SST3 could play a signi cant pathophysio- fi logic role in NFPT cells, we rst corroborated that SST3 was the highest SST3-agonist reduced tumor growth in an in vivo preclinical overexpressed receptor in our NFPT cohort, and demonstrated that model SST3 mRNA levels could discriminate between NFPTs and function- To demonstrate the ability of SST3 to mediate a reduction in ing PitNETs, but not between NPs and functioning PitNETs. More- in vivo tumor growth , we used the POMC-KO mice, a mouse model over, use of IHC also validated the presence of SST3 protein in of PitNETs. Indeed, IHC analysis of the tumors formed confirmed approximately 50% of NFPTs analyzed, using two different cohorts the presence of SST3-staining in all tumors (Fig. 6A), and a strong of patients, which presented both a membrane and cytoplasmic presence of Ki67-staining in the tumor area in contrast to the localization, as described previously (9). To further explore the func- normal/adjacent mice pituitary gland (Fig. 6B). Mice were treated tional relevance of this receptor in NFPTs, we compared the effects of

AB Human NFPT Mice PA Tumor

Normal pituitary

Tumor

C D 30 Treatment Vehicle n = 16 ) 80 25.206 3 BIM-355 2.2 mg/kg/day n = 8 25 P = 0.09 20.124 P = 0.02 60 BIM-355 7.4 mg/kg/day n = 3 20 17.069 * Indicates significantly 40 15 different from vehicle 10 20 5 Geomtric LS means Tumor volume (mm volume Tumor 0 0 -4 -20 2 4 6 8 Pump SC Pump7day + 355 Pump7day + 355 Treatment week vehicle 2 mg/kg/day high 7.4 mg/kg/day

Figure 6.

Direct effect of BIM-355 on tumor growth in a preclinical PitNET mouse model. A, Representative IHC of SST3 in tumors generated in this model compared with SST3 protein expression in human NFPTs (200 magnification; scale bar, 20 mm). B, Representative IHC of Ki67 in normal adjacent pituitary gland and tumors formed in this model (200 magnification; scale bar, 20 mm). C, Effect of BIM-355 [2.2 mg/kg/day (n ¼ 8) and 7.4 mg/kg/day (n ¼ 3)] on tumor volume growth during 8 weeks, measured by MRI. The asterisks represent statistical significance evaluated time-by-time by geometric least square values adjusted for baseline tumor volume. D, Final tumor volume after 8 weeks of treatment. The bars represent geometric least square mean values corresponding to the complete curves from baseline to week 8, as described in the Statistical Analysis section.

OF10 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Role of SST3 in Nonfunctioning Pituitary Tumors

fi the two peptidic agonists identi ed herein with those exerted by the supported by ROC curve analysis, demonstrating that only SST3 levels, previously published nonpeptidic agonist L-796,778 (20). The first but not that of the other SST subtypes, were able to discriminate approach was to assess the effect of SST3-agonists treatment on cell between responsive and unresponsive NFPTs. SST3-agonists were able viability, a parameter tightly linked to tumor growth, the main clinical to increase SST3 expression levels in responsive, but not in unrespon- problem associated to NFPTs. Our data revealed a general ability of sive NFPTs (the latter showing even a reduction of SST3 expression), SST3-agonists to reduce cell viability in NFPT cells cultured in vitro, thereby suggesting that a positive feedback, and thus a functionally SSTR3 wherein BIM-355 was the most potent agonist (i.e., 23.2% of reduc- unique SST3(receptor)- (gene) regulatory circuit might operate tion), which, to our knowledge, is the most prominent reduction in these tumors. Remarkably, this is neither the sole nor the first time reached to date in a cell culture model on cell viability through the that an up-regulation of SSTs expression levels has been reported in activation of SST3 [e.g., treatment with the SST3-agonist (L-796,778) response to different agonists. Thus, we have previously reported an exerted a 19% inhibition on cell proliferation in HEK-293 cells in increase of SST2 and SST5 expression levels in response to the chimeric previous studies (18)]. At this point, it is important to note that compound BIM-23A760 in pituitary cells (22). Nevertheless, the fi pituitary tumor cells, and specially NFPT cells, are not as markedly speci c reasons why SST3 reduction only occurs in unresponsive responsive to pharmacologic treatments as cells from other tumor NFPTs need to be further explored. types (7, 22). Most importantly, we also observed a clear increase of To interrogate the mechanisms underlying the response to treat- caspase activity, which is likely translated into an increase of apoptosis, ment with SST3-agonists, we explored an ample range of signaling in response to treatment with different SST3-agonists. These effects on pathways. In contrast with the scarce association (if any) observed 2þ caspase activity, as well as those on cell viability, were completely between [Ca ]i dynamics and response to SST3-agonists, our data blocked by the antagonist BIM-839, thus confirming the specificity of revealed, for the first time, a striking reduction on the phosphorylation the actions observed. These results are in line with the previous results levels of relevant protein kinases associated to three important sig- mentioned above, which indicated that treatment with L-796,778 naling pathways (i.e., MAPK, PI3K-AKT/mTOR, and JAK/STAT). increased the apoptotic index or related markers in the HEK-293 cell Specifically, we observed a decrease on ERK1/2 and JNK phosphor- line (18). Moreover, similar effects were reported in SST3-overexpres- ylation, as well as in other proteins of the MAPK pathway such as sing breast cancer cells (33), further suggesting a clear, robust anti- RSK1/2/3, CREB, and c-Jun in response to BIM-355 in responsive proliferative and proapoptotic role of this receptor. Of note, in support NFPTs, but not in unresponsive NFPTs. In fact, these pathways have of the antiproliferative and proapoptotic role of SST3 in NFPTs been commonly associated with cell growth, proliferation, and cell observed in the current study, we demonstrated, for the first time, survival in tumor pathologies, including NFPTs (39–41). Moreover, in vivo fi fi that administration of an SST3-speci c agonist signi cantly our results demonstrating the importance of the MAPK pathway in the decreased tumor growth in a preclinical mouse model of Pit- response to SST3-agonists in NFPTs are in line with previous results on NET (27, 28). It has to be also noted that the differential response HEK-293 cells treated with L-796,778 (18). Similarly, we also found a observed in the functional assays depending on the SST3-agonist used decreased phosphorylated status in Akt, GSK-3a/b, Src, PRAS40, may be due to the highly complex activation of G protein–coupled mTOR, and p70-S6 kinase, all of them associated to the PI3K/ receptors (GPCR), including the presence of multiple ligand-binding AKT/mTOR pathway, which controls cell-cycle progression, protein sites that can influence signal transduction in a distinct manner (34). synthesis, and cell proliferation (41). Indeed, it has been demonstrated Our data also revealed a reduction of CGA/FSH/LH mRNA found in that SSAs, such as octreotide, elicit their antiproliferative actions response to BIM-355 (the most effective compound), which is not rare through inhibition of PI3K/AKT pathway (42), and also that specific because we have previously observed a reduction of CGA and other inhibitors of mTOR pathway exhibit a potent antitumor efficacy in hormones at mRNA levels in response to other treatments in Pit- NFPTs (43). Remarkably, all the changes in phosphorylation levels are NETs (22). In addition, we found a decrease of CgA secretion in NFPT consistent with reduced PI3K/AKT/mTOR signaling, except for p53, in vitro cells cultured in response to SST3-agonists. CgA has been an important tumor suppressor, whose phosphorylation is involved in detected in secretory granules of NFPT cells, its secretion has been apoptosis. In fact, our results show a reduction on TP53 mRNA and employed as a useful marker of pharmacological effectiveness, and has protein expression and phosphorylated-p53 levels, which is in line been recently proposed as a biomarker of tumorigenesis and inva- with the reduction observed in C6 glioma cell in response to acet- siveness since it was altered in invasive NFPTs (31, 35, 36), again aminophen (44). It is well established that p53 is regulated by complex supporting SST3 as an attractive therapeutic target in NFPTs. In this mechanisms. Nevertheless, the reduced levels of p53 remain to be sense, it is important to note that the development of the nonpeptidic confirmed. In addition, we observed a reduction on the phosphory- SST3 agonist L-796,778 has been stopped and, therefore, the new lation levels of some components of JAK/STAT signaling pathway peptides described in this work open a new avenue for the future such as STAT3, STAT5a/b, and STAT6. In marked contrast with development and use of SST3-agonists to treat these tumors. responsive tumors, in unresponsive NFPTs, the phosphorylation Interestingly, we found that a subset of NFPTs did not respond status of most of the proteins described above were not altered upon to SST3-agonists in terms of inhibition of cell viability. This is not BIM-355 treatment, or were increased. This situation is in accordance surprising, in that previous studies have shown a percentage with results showing an increase on phosphorylation of Akt levels in a of unresponsive NFPTs and functioning PitNETs, or even total proportion of NFPTs resistant to rapamycin (45) and recent results absence of effect, upon treatment with different antitumor com- published in prolactinomas showing that a reduced D2S/D2L expres- pounds (7, 37, 38). A comparison of the expression profile of SSTs sion ratio and an increase on MAPK and PI3K/AKT/mTOR pathways revealed higher SST3 expression levels in responsive versus unrespon- might contribute to tumorigenesis and DA resistance (46). Moreover, sive tumors, while no other significant changes were found between the comparing the basal phosphorylation levels of vehicle-treated controls two populations. Moreover, this result was corroborated at protein from responsive and unresponsive NFPTs revealed a clear pattern with levels (higher SST3 in responsive vs. unresponsive tumors), thus lower basal phosphorylation levels of the MAPK, PI3K-AKT/mTOR, fi suggesting that high SST3 levels are a requisite to elicit a signi cant and JAK/STAT proteins in unresponsive tumors, which could explain response after SST3-agonists administration. This notion is further the resistance of this population of NFPTs to respond to SST3-agonists

AACRJournals.org Clin Cancer Res; 2020 OF11

Vazquez-Borrego et al.

(and possibly to other SSAs). Nevertheless, the interpretation of these functional role in the pathophysiology of NFPTs, and invites to suggest results should be taken with caution because these results are derived that pharmacologic treatments specifically targeting this receptor from a pool of three responsive and three unresponsive NFPTs. could become a promising option to treat patients with NFPTs, Consistent with the existence of two populations of NFPTs (at least providing a relevant clinical conclusion, which should be soon tested in terms of reduction of cell viability and signaling capacity in response for their use in humans. to SST3-agonists), the experimental silencing of SSTR3 gene expression by specific siRNAs increased cell viability in 38% of NFPTs ana- Disclosure of Potential Conflicts of Interest lyzed. A potential explanation for this observation would be that J. Hill, H. Halem, and M. Paez-Pereda are employees/paid consultants for Ipsen Biosciences Inc. J.P. Castano~ reports receiving commercial research grants from SST3 bears a constitutive inhibitory activity, as previous results have demonstrated that various SSTs, including SST , display a relevant Novartis and Ipsen, speakers bureau honoraria from Prime, Ipsen, and Novartis, and 3 holds ownership interest (including patents) in University of Cordoba. No potential degree of ligand-independent constitutive activity in different pitu- conflicts of interest were disclosed by the other authors. itary cell systems (47, 48). Another possibility would be the endogenous release of somatostatin or its relative peptide cortistatin. In Authors’ Contributions this sense, our results revealed almost negligible expression levels of Conception and design: M.C. Vazquez-Borrego, V. Gupta, A. Ibanez-Costa,~ somatostatin and much higher levels of cortistatin, which might J.P. Castano,~ R.M. Luque suggest the existence of an ultra-short autocrine feedback mecha- Development of methodology: M.C. Vazquez-Borrego, V. Gupta, A. Ibanez-Costa,~ H. Halem, J.P. Castano,~ R.M. Luque nism whereby activation of SST3 by endogenous cortistatin would Acquisition of data (provided animals, acquired and managed patients, provided mediate inhibition of cell viability, secretion, etc. Inversely, a reduction in SST protein levels (by silencing) could lead to the facilities, etc.): E. Venegas-Moreno, A. Toledano-Delgado, D.A. Cano, R. Ortega- 3 Salas, M.A. Japon, A. Barrera-Martín, A. Vasiljevic, S. Zhang, H. Halem, J. Solivera, loss of this inhibitory loop and could, ultimately, induce an increase G. Raverot, M.A. Galvez, A. Soto-Moreno, R.M. Luque on cell viability. However, the levels of somatostatin and cortistatin Analysis and interpretation of data (e.g., statistical analysis, biostatistics, were not different between SSTR3 silencing responsive and unre- computational analysis): M.C. Vazquez-Borrego, A. Ibanez-Costa,~ M.D. Gahete, sponsive tumors, and this hypothesis should be further explored. S. Zhang, H. Halem, M. Paez-Pereda, M.D. Culler, J.P. Castano,~ R.M. Luque Importantly, the fact that BIM-355 did not alter cell viability in Writing, review, and/or revision of the manuscript: M.C. Vazquez-Borrego, ~ s13501-transfected cells (with reduced SST content), together with V. Gupta, A. Ibanez-Costa, M.D. Gahete, E. Venegas-Moreno, D.A. Cano, 3 C. Blanco-Acevedo, R. Ortega-Salas, M.A. Japon, A. Barrera-Martín, A. Vasiljevic, the fact that SST3-antagonist BIM-839 completely blocked the S. Zhang, H. Halem, J. Solivera, G. Raverot, M.A. Galvez, A. Soto-Moreno, M. Paez- effects of SST3-agonists, provide compelling evidence that, although Pereda, M.D. Culler, J.P. Castano,~ R.M. Luque cAMP IC50 for BIM-355 for SST3 and SST5 are very similar, the Administrative, technical, or material support (i.e., reporting or organizing data, effects observed in response to agonists are completely and only constructing databases): M.C. Vazquez-Borrego, C. Blanco-Acevedo, J.P. Castano,~ specific of SST . These observations unveiled new conceptual and R.M. Luque 3 ~ functional avenues in NFPTs, with potential therapeutic implica- Study supervision: M.C. Vazquez-Borrego, M. Paez-Pereda, J.P. Castano, M.D.Gahete,R.M.Luque tions, which are worth exploring in future studies. Other (synthesis and supply of compounds for testing; data analysis): J. Hill Taken together, this study represents the first identification and characterization of SST3 peptidic agonists, which enabled to gain novel Acknowledgments experimental results supporting that these agonists exert clear anti- This work has been funded by the following grants: Junta de Andalucía [CTS-1406 tumor actions on NFPT primary cell cultures. Moreover, our data (R.M. Luque), BIO-0139 (J.P. Castano)];~ Ministerio de Ciencia, Innovacion y Uni- versidades [BFU2016-80360-R (J.P. Castano)]~ and Instituto de Salud Carlos III, co- reveal a role of SST3 in key pathophysiologic processes of NFPT cells funded by European Union [ERDF/ESF, “Investing in your future”: PI16/00264 (R.M. such as cell viability, apoptosis, and CgA secretion when SST3 is Luque), CP15/00156 (M.D. Gahete) and CIBERobn]. CIBER is an initiative of targeted with specific agonists. These actions are likely mediated Instituto de Salud Carlos III. through the modulation of three important signaling pathways (MAPK, PI3K-AKT/mTOR, and JAK/STAT). We also found two The costs of publication of this article were defrayed in part by the payment of page populations of NFPTs with a differential expression of SST3 (statis- charges. This article must therefore be hereby marked advertisement in accordance tically significant at the mRNA and protein level) and different with 18 U.S.C. Section 1734 solely to indicate this fact. capacity to respond to SST3-agonists in terms of activation of signaling pathways and functional parameters. Therefore, the current study Received June 30, 2019; revised September 8, 2019; accepted October 14, 2019; fi provides new compelling evidence, demonstrating that SST3 has a published rst October 17, 2019.

References 1. Theodoropoulou M, Stalla GK. Somatostatin receptors: from signaling to clinical gastroenteropancreatic neuroendocrine tumors. Endocrine 2018;59: practice. Front Neuroendocrinol 2013;34:228–52. 426–37. 2. Ben-Shlomo A, Melmed S. Pituitary somatostatin receptor signaling. 6. Elston MS, Meyer-Rochow GY, Conaglen HM, Clarkson A, Clifton-Bligh RJ, Trends Endocrinol Metab 2010;21:123–33. Conaglen JV, et al. Increased SSTR2A and SSTR3 expression in succinate 3. Vitale G, Dicitore A, Sciammarella C, Di Molfetta S, Rubino M, Faggiano A, et al. dehydrogenase-deﬁcient pheochromocytomas and paragangliomas. Hum Pasireotide in the treatment of neuroendocrine tumors: a review of the literature. Pathol 2015;46:390–6. Endocr Relat Cancer 2018;25:R351–R64. 7. Ibanez-Costa A, Rivero-Cortes E, Vazquez-Borrego MC, Gahete MD, 4. Melmed S, Colao A, Barkan A, Molitch M, Grossman AB, Kleinberg D, et al. Jimenez-Reina L, Venegas-Moreno E, et al. Octreotide and pasireotide Guidelines for acromegaly management: an update. J Clin Endocrinol Metab (dis)similarly inhibit pituitary tumor cells in vitro. J Endocrinol 2016; 2009;94:1509–17. 231:135–45. 5. Herrera-Martinez AD, Gahete MD, Pedraza-Arevalo S, Sanchez-San- 8. Hoﬂand LJ, Feelders RA, de Herder WW, Lamberts SW. Pituitary tumours: chez R, Ortega-Salas R, Serrano-Blanch R, et al. Clinical and func- the sst/D2 receptors as molecular targets. Mol Cell Endocrinol 2010;326: tional implication of the components of somatostatin system in 89–98.

OF12 Clin Cancer Res; 2020 CLINICAL CANCER RESEARCH

Role of SST3 in Nonfunctioning Pituitary Tumors

9. Lee M, Lupp A, Mendoza N, Martin N, Beschorner R, Honegger J, et al. SSTR3 is 30. Heaney AP, Curry WJ, Pogue KM, Armstrong VL, Mirakhur M, Sheridan B, et al. a putative target for the medical treatment of gonadotroph adenomas of the Immunohistochemical evaluation of the post-translational processing of chro- pituitary. Endocr Relat Cancer 2015;22:111–9. mogranin A in human pituitary adenomas. Pituitary 2000;3:67–75. 10. Asa SL, Ezzat S. The pathogenesis of pituitary tumors. Annu Rev Pathol 2009;4: 31. Zatelli MC, Piccin D, Bottoni A, Ambrosio MR, Margutti A, Padovani R, et al. 97–126. Evidence for differential effects of selective somatostatin receptor subtype 11. Greenman Y, Stern N. Optimal management of non-functioning pituitary agonists on alpha-subunit and chromogranin a secretion and on cell viability adenomas. Endocrine 2015;50:51–5. in human nonfunctioning pituitary adenomas in vitro. J Clin Endocrinol Metab 12. Greenman Y, Stern N. Non-functioning pituitary adenomas. Best Pract Res Clin 2004;89:5181–8. Endocrinol Metab 2009;23:625–38. 32. Gruppetta M, Formosa R, Falzon S, Ariff Scicluna S, Falzon E, Degeatano J, et al. 13. Oystese KA, Evang JA, Bollerslev J. Non-functioning pituitary adenomas: growth Expression of cell cycle regulators and biomarkers of proliferation and regrowth and aggressiveness. Endocrine 2016;53:28–34. in human pituitary adenomas. Pituitary 2017;20:358–71. 14. Cortet-Rudelli C, Bonneville JF, Borson-Chazot F, Clavier L, Coche Dequeant B, 33. War SA, Kim B, Kumar U. Human somatostatin receptor-3 distinctively induces Desailloud R, et al. Post-surgical management of non-functioning pituitary apoptosis in MCF-7 and cell cycle arrest in MDA-MB-231 breast cancer cells. adenoma. Ann Endocrinol 2015;76:228–38. Mol Cell Endocrinol 2015;413:129–44. 15. Kanner AA, Corn BW, Greenman Y. Radiotherapy of nonfunctioning and 34. Latorraca NR, Venkatakrishnan AJ, Dror RO. GPCR dynamics: structures in gonadotroph adenomas. Pituitary 2009;12:15–22. motion. Chem Rev 2017;117:139–55. 16. Raverot G, Vasiljevic A, Jouanneau E. Prognostic factors of regrowth in 35. Yu SY, Hong LC, Feng J, Wu YT, Zhang YZ. Integrative proteomics and nonfunctioning pituitary tumors. Pituitary 2018;21:176–82. transcriptomics identify novel invasive-related biomarkers of non-functioning 17. Kopczak A, Renner U, Karl Stalla G. Advances in understanding pituitary pituitary adenomas. Tumour Biol 2016;37:8923–30. tumors. F1000Prime Rep 2014;6:5. 36. Riva C, Leutner M, Capella C, Usellini L, la Rosa S, Brianza MR, et al. Different 18. War SA, Somvanshi RK, Kumar U. Somatostatin receptor-3 mediated intracel- expression of chromogranin A and chromogranin B in various types of pituitary lular signaling and apoptosis is regulated by its cytoplasmic terminal. adenomas. Zentralbl Pathol 1993;139:165–70. Biochim Biophys Acta 2011;1813:390–402. 37. Florio T, Barbieri F, Spaziante R, Zona G, Hofland LJ, van Koetsveld PM, 19. Sharma K, Patel YC, Srikant CB. Subtype-selective induction of wild-type p53 et al. Efficacy of a dopamine-somatostatin chimeric molecule, BIM-23A760, and apoptosis, but not cell cycle arrest, by human somatostatin receptor 3. in the control of cell growth from primary cultures of human non- Mol Endocrinol 1996;10:1688–96. functioning pituitary adenomas: a multi-center study. Endocr Relat Cancer 20. Rohrer SP, Birzin ET, Mosley RT, Berk SC, Hutchins SM, Shen DM, et al. Rapid 2008;15:583–96. identification of subtype-selective agonists of the somatostatin receptor through 38. Cuevas-Ramos D, Fleseriu M. Somatostatin receptor ligands and resistance to combinatorial chemistry. Science 1998;282:737–40. treatment in pituitary adenomas. J Mol Endocrinol 2014;52:R223–40. 21. Luque RM, Ibanez-Costa~ A, Lopez-S anchez LM, Jimenez-Reina L, Venegas- 39. Rubinfeld H, Shimon I. PI3K/Akt/mTOR and Raf/MEK/ERK signaling path- Moreno E, Galvez MA, et al. A cellular and molecular basis for the selective ways perturbations in non-functioning pituitary adenomas. Endocrine 2012;42: desmopressin-induced ACTH release in cushing disease patients: key role of 285–91. AVPR1b receptor and potential therapeutic implications. J Clin Endocrinol 40. Xiao X, Li BX, Mitton B, Ikeda A, Sakamoto KM. Targeting CREB for cancer Metab 2013;98:4160–9. therapy: friend or foe. Curr Cancer Drug Targets 2010;10:384–91. 22. Ibanez-Costa A, Lopez-Sanchez LM, Gahete MD, Rivero-Cortes E, Vazquez- 41. Roof AK, Gutierrez-Hartmann A. Consider the context: Ras/ERK and PI3K/ Borrego MC, Galvez MA, et al. BIM-23A760 influences key functional endpoints AKT/mTOR signaling outcomes are pituitary cell type-specific. Mol Cell Endo- in pituitary adenomas and normal pituitaries: molecular mechanisms underlying crinol 2018;463:87–96. the differential response in adenomas. Sci Rep 2017;7:42002. 42. Theodoropoulou M, Zhang J, Laupheimer S, Paez-Pereda M, Erneux C, 23. Alhambra-Exposito MR, Ibanez-Costa A, Moreno-Moreno P, Rivero-Cortes E, Florio T, et al. Octreotide, a somatostatin analogue, mediates its antiproli- Vazquez-Borrego MC, Blanco-Acevedo C, et al. Association between radiolog- ferative action in pituitary tumor cells by altering phosphatidylinositol ical parameters and clinical and molecular characteristics in human somato- 3-kinase signaling and inducing Zac1 expression. Cancer Res 2006;66: tropinomas. Sci Rep 2018;8:6173. 1576–82. 24. Xie F, Xiao P, Chen D, Xu L, Zhang B. miRDeepFinder: a miRNA analysis tool for 43. LeeM,WiedemannT,GrossC,LeinhauserI,RoncaroliF,BrarenR,etal. deep sequencing of plant small RNAs. Plant Mol Biol 2012 Jan 31. [Epub ahead of Targeting PI3K/mTOR signaling displays potent antitumor efficacy print]. against nonfunctioning pituitary adenomas. Clin Cancer Res 2015;21: 25. Hormaechea-Agulla D, Gahete MD, Jimenez-Vacas JM, Gomez-Gomez E, 3204–15. Ibanez-Costa A, F LL, et al. The oncogenic role of the In1-ghrelin splicing 44. Lee YS, Wan J, Kim BJ, Bae MA, Song BJ. Ubiquitin-dependent degradation of variant in prostate cancer aggressiveness. Mol Cancer 2017;16:146. p53 protein despite phosphorylation at its N terminus by acetaminophen. 26. Ibanez-Costa~ A, Gahete MD, Rivero-Cortes E, Rincon-Fernandez D, Nelson R, J Pharmacol Exp Ther 2006;317:202–8. Beltran M, et al. In1-ghrelin splicing variant is overexpressed in pituitary 45. Cerovac V, Monteserin-Garcia J, Rubinfeld H, Buchfelder M, Losa M, Florio T, adenomas and increases their aggressive features. Sci Rep 2015;5:8714. et al. The somatostatin analogue octreotide confers sensitivity to rapamycin 27. Karpac J, Ostwald D, Li GY, Bui S, Hunnewell P, Brennan MB, et al. Proopio- treatment on pituitary tumor cells. Cancer Res 2010;70:666–74. melanocortin heterozygous and homozygous null mutant mice develop pituitary 46. Roof AK, Jirawatnotai S, Trudeau T, Kuzyk C, Wierman ME, Kiyokawa H, et al. adenomas. Cell Mol Biol 2006;52:47–52. The balance of PI3K and ERK signaling is dysregulated in prolactinoma and 28. Yaswen L, Diehl N, Brennan MB, Hochgeschwender U. Obesity in the mouse modulated by dopamine. Endocrinology 2018;159:2421–34. model of pro-opiomelanocortin deficiency responds to peripheral melanocortin. 47. Eigler T, Ben-Shlomo A, Zhou C, Khalafi R, Ren SG, Melmed S. Constitutive Nat Med 1999;5:1066–70. somatostatin receptor subtype-3 signaling suppresses growth hormone synthe- 29. Lupp A, Nagel F, Doll C, Rocken C, Evert M, Mawrin C, et al. Reassessment of sis. Mol Endocrinol 2014;28:554–64. sst3 somatostatin receptor expression in human normal and neoplastic tissues 48. Ben-Shlomo A, Pichurin O, Barshop NJ, Wawrowsky KA, Taylor J, Culler MD, using the novel rabbit monoclonal antibody UMB-5. Neuroendocrinology 2012; et al. Selective regulation of somatostatin receptor subtype signaling: evidence 96:301–10. for constitutive receptor activation. Mol Endocrinol 2007;21:2565–78.

AACRJournals.org Clin Cancer Res; 2020 OF13

A Somatostatin Receptor Subtype-3 (SST3) Peptide Agonist Shows Antitumor Effects in Experimental Models of Nonfunctioning Pituitary Tumors

Mari C. Vázquez-Borrego, Vandana Gupta, Alejandro Ibáñez-Costa, et al.

Clin Cancer Res Published OnlineFirst October 17, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-19-2154

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2019/10/17/1078-0432.CCR-19-2154.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2020/01/03/1078-0432.CCR-19-2154. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.